On Wed, Jan 29, 2014 at 6:28 AM, Dan Fairs <[email protected]> wrote:
> Hi,
>
> I'm using PostgreSQL 9.3.2 with WAL-E 0.6.6.
>
> I'm experimenting with WAL-E PITR recovery. I was successful with a few
> small databases, and now I'm trying a production-size database (about 76GB).
>
> Base backup and subsequent log archival go well. I'm also able to restore
> the base backup onto a slave. When I then place a recovery.conf in the
> slave's cluster data directory and start the server, it starts fetching WAL
> from the archive using WAL-E but then terminates with a message like this:

I am tracking a problem with backup-fetch that is rather scary.   Can
you send to me -- privately, or on list -- a redacted transcript of
your backup-fetch logs?

So far the bug has been extremely rare in practice, even though it's
been latent for years.  I committed what I think is a fix, but haven't
released it yet because I haven't been able to get a neat and tidy
reproduction under simulation.

I may wind up releasing anyway, because there is some amount of
obvious wrongness-by-inspection and I don't want to sit on my hands
for too long.

(patch, for the curious:
https://github.com/wal-e/wal-e/commit/19c204f60ab493492ba84de0b5040d5efccc9f95)

-- 
You received this message because you are subscribed to the Google Groups 
"wal-e" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to